Prevent gaps in tall delimiters #1986
Conversation
|
@kevinbarabash This PR does several vertical alignment calculations and the screenshotter tests declare several errors. It appears to me that all of them are sub-pixel issues except the Array test, in which the diff shows one thin green line. I'm willing to declare this PR as good to go, but I'll leave the screenshotter tests as is for now. If you agree that the PR is okay, I'll replace the screenshots with new ones. |
|
I went ahead and updated the screenshots. |
Codecov Report
@@ Coverage Diff @@
## master #1986 +/- ##
==========================================
+ Coverage 93.36% 93.38% +0.01%
==========================================
Files 79 79
Lines 4931 4942 +11
Branches 860 860
==========================================
+ Hits 4604 4615 +11
Misses 287 287
Partials 40 40
Continue to review full report at Codecov.
|
1 similar comment
Codecov Report
@@ Coverage Diff @@
## master #1986 +/- ##
==========================================
+ Coverage 93.36% 93.38% +0.01%
==========================================
Files 79 79
Lines 4931 4942 +11
Branches 860 860
==========================================
+ Hits 4604 4615 +11
Misses 287 287
Partials 40 40
Continue to review full report at Codecov.
|
| // at a position that juts 0.005 above the bottom of the top element. | ||
| inners.push({type: "kern", size: shiftOfExtraElement}); | ||
| inners.push(makeInner(repeat, font, mode)); | ||
| inners.push(lap); |
There was a problem hiding this comment.
In the I see, there's two else branch we do the same thing. Do we need to insert two extra repeat elements?for loops in the else block.
Fixes #1975 by typesetting tall delimiters with a small (0.005em) overlap where glyphs meet.
As evidence, I offer two screenshots. The first one is before this PR:
The second screenshot is after this PR:
The value of 0.005em is entirely a guess on my part and may need later adjustment.